Processor unit and method for predictive speed control of a motor vehicle

Abstract

The invention relates to a processor unit (1) for the predictive control of a motor vehicle (2), the processor unit (1) having a memory unit (4) with an MPC (Model Predictive Control) module (5) which contains a cost function (J), the cost function (J) having a time term (Z) weighted by a first weighting factor (# imgabs0 #), the cost function (J) has a time term which contains a driving time, which is predicted independently of a drive train variant of the motor vehicle (2) and is required for the motor vehicle (2) to pass through a mileage predicted throughout the prediction field of vision, and wherein the cost function (J) has an efficiency term (E) weighted with a second weighting factor (# imgabs1 #), the processor unit (1) is configured to execute an MCP module (5), which provides an energy consumption predicted within a predicted field of view as a function of at least a vehicle speed (# imgabs2 #) and a wheel force (# imgabs3 #), the provided efficiency item (E) being independent of a drive train variant of the motor vehicle (2), the processor unit (1) being configured to execute the MCP module (5) as a function of a weighted efficiency item (E) and a weighted time item (Z), and to execute the MCP module (5) as a function of the weighted efficiency item (E) and the weighted time item (Z). A speed control for the motor vehicle (2) with respect to the predicted field of vision is determined for the motor vehicle (2) independently of a drive train variant of the motor vehicle (2), such that the cost function (J) is minimized. Furthermore, the invention relates to a method and a computer program product.

Description

Technical Field

[0001] The invention relates to a processor unit and a method for predictive speed control of a motor vehicle. The invention also relates to a computer program product. Background Art

[0002] When a motor vehicle is in operation, the driver and his or her driving style have a significant impact on energy consumption. Cruise control systems may only take the route topology into account.

[0003] Therefore, in particular in the area of ​​engine control in motor vehicles, methods based on model-predictive control (English: Model Predictive Control, abbreviated to MPC) can be used for control.

[0004] These optimization-based controls, in addition to requiring a high level of computing power and computing time, also require precise knowledge of the motor vehicle and the drive train, so that their use for online computing is problematic on the one hand and also requires a high level of development effort.

[0005] In his paper, "Energy-Efficient Driver Assistance System for Electric Vehicles Using Model-Predictive Control," (Schwickart T., PhD thesis, University of Luxembourg, 2015), Schwickart proposed a quadratic programming-based approach. This approach transforms the nonlinear system model into a linear one, resulting in a quadratic optimization problem that is easily solved numerically.

[0006] Therefore, DE 10 2020 216251 A1 discloses a processor unit for model-predictive control of a motor vehicle, wherein the processor unit is configured to: call route topography data, the route topography data including information about the topography of a road section ahead of the motor vehicle on which the motor vehicle should travel in order to reach a predetermined mileage point; call traffic condition data, the traffic condition data including information about traffic events on the road section ahead of the motor vehicle; call driving characteristics data, the driving characteristics data including information about criteria according to which the operation of multiple units of the motor vehicle on the road section ahead of the motor vehicle should be optimized; implement an MPC algorithm for model-predictive control of the motor vehicle, wherein the MPC algorithm includes a longitudinal dynamics model of the motor vehicle and at least one cost function to be minimized; determine the speed trajectory of the motor vehicle and the operating strategy of the units of the motor vehicle for a future sliding prediction horizon by implementing the MPC algorithm while taking into account the longitudinal dynamics model of the motor vehicle, so as to minimize at least one cost function, wherein the operating strategy is determined based on the route topography data, the traffic condition data and the driving characteristics data.

[0007] DE 10 2020203742 A1 discloses a processor unit for model-predictive control of a motor vehicle, wherein the processor unit is configured to: implement an MPC algorithm, the MPC algorithm comprising a first solution module, a second solution module, a third solution module, a longitudinal dynamics model, and three cost functions, wherein a first cost function is assigned to the first solution module, wherein a second cost function is assigned to the second solution module, and wherein a third cost function is assigned to the third solution module; calculate a speed trajectory that minimizes the first cost function by implementing the first solution module for a road section ahead while taking into account the longitudinal dynamics model, the motor vehicle in the predicted The vehicle should proceed along the speed trajectory within the prediction horizon; and calculate a curve of the battery's state of charge that minimizes a first cost function, the battery being used as an energy storage device for the motor of the motor vehicle; implement a second solution module based on the speed trajectory and the battery's state of charge curve and calculate a trajectory of an integer control parameter that minimizes the second cost function while taking into account constraints; and implement a third solution module for the starting section of the prediction horizon to calculate a torque trajectory that minimizes a third cost function for the motor, internal combustion engine and braking device of the motor vehicle, and the motor, internal combustion engine and braking device should provide torque within the prediction horizon according to the torque trajectory.

[0008] DE102019216445A1 discloses a processor unit for model-predictive control of an electric motor of a drive train of a motor vehicle, wherein the processor unit is configured to: implement an MPC algorithm for model-predictive control of an electric motor of a drive train of a motor vehicle, the MPC algorithm including a longitudinal dynamics model of the drive train, the MPC algorithm also including a cost function to be minimized, the cost function including as a first term the electrical energy provided by a battery of the drive train to drive the electric motor within a prediction horizon, weighted by a first weighting factor and predicted according to the longitudinal dynamics model, the cost function also including as a second term the driving time required for the motor vehicle to travel the entire mileage predicted within the prediction horizon, weighted by a second weighting factor and predicted according to the longitudinal dynamics model, and the processor unit is configured to: by implementing the MPC algorithm in dependence on the first term and in dependence on the second term, obtain input variables for the electric motor so as to minimize the cost function. Summary of the Invention

[0009] Therefore, the object of the present invention is to provide an improved processor unit and an improved method for predictive control of a motor vehicle. Furthermore, the object is to provide a computer program product.

[0010] This object is achieved by a processor unit having the features of claim 1 , a method having the features of claim 10 , and a computer program product having the features of claim 14 .

[0011] The object is achieved by a processor unit for predictive control of a motor vehicle, wherein the processor unit has a memory unit with an MPC (Model Predictive Control) module, the MPC module including a cost function, wherein the cost function has a time term weighted by a first weighting factor, the time term including a travel time required by the motor vehicle for traveling the entire distance predicted within a prediction horizon, independently of a drive train variant of the motor vehicle, and wherein the cost function has an efficiency term weighted by a second weighting factor, which efficiency term provides an energy consumption prediction within a prediction horizon that is dependent at least on the vehicle speed and the wheel forces, wherein the provided efficiency term is independent of a drive train variant of the motor vehicle, Therein, the processor unit is configured to: determine a speed control for the motor vehicle regarding a prediction horizon for the motor vehicle independently of the drive system variant of the motor vehicle by implementing the MCP module in dependence on a weighted efficiency term and a weighted time term, so as to minimize a cost function.

[0012] According to the present invention, it has been recognized that all processor units described in the prior art for MPC-based planning are based on specific drivetrain variants. In other words, each processor unit is always tailored to the respective drivetrain, such as for a purely internal combustion engine vehicle, a hybrid vehicle, or a purely electric vehicle, such as an electric vehicle with a single electric machine and a single transmission stage. Therefore, it is currently not possible to transfer the respective MPC modules to other drivetrain variants.

[0013] The present invention furthermore recognized that developing and coordinating a separate MPC module for each drivetrain variant results in unacceptably high application complexity. The problem to be solved is compounded by the large number of possible solutions, such as hybrid vehicles with electric drives, fuel cells, purely electric drives with varying numbers of motors and transmissions, or purely internal combustion engines. Continuous developments also require continuous updating. Furthermore, covering all possible variants is virtually impossible, particularly in terms of time and resulting development costs.

[0014] Furthermore, it has been recognized that the coordination effort increases disproportionately due to the larger number of optimization variables and cost function terms.

[0015] Furthermore, the present invention has recognized that artificially expanding the representation of a superset of variables that may exist in different drivetrain variants is detrimental to the real-time capabilities of the processor unit. Furthermore, experiments have shown that these artificial expansions can lead to undesirable planning results under certain circumstances, which in turn require situation-dependent optimization restrictions or optimization boundary conditions, which are difficult to handle.

[0016] Based on these findings according to the present invention, a generally effective formulation of the MPC problem was deliberately developed with the goal of developing a processor unit that is efficient / usable for all drive train variants.

[0017] This is achieved by means of a processor unit according to the present invention for predictive control of a motor vehicle. The processor unit according to the present invention provides a general description of efficiency using the efficiency term according to the present invention, which uniformly provides the current energy consumption over time for a wide range of drivetrain variants with respect to the prediction horizon.

[0018] The cost function includes as a time term the travel time required by the motor vehicle to cover the entire distance predicted within the prediction horizon, weighted by a first weighting factor and predicted, in particular, according to a generally applicable longitudinal dynamics model. The time term in the cost function means that, depending on the selection of the weighting factor, lower speeds are not always evaluated as optimal in terms of efficiency, and thus the problem of the resulting speed always being at the lower limit of the permissible speed range no longer exists.

[0019] The cost function also includes an efficiency term that is independent of a specific drive train and that predicts the energy required to drive the motor vehicle within a prediction horizon, independent of a specific energy source such as batteries or fuel / gasoline. The processor unit is configured to determine a speed control for the motor vehicle by implementing an MPC module that is dependent on the time term and the efficiency term, such that the cost function is minimized.

[0020] The efficiency term is based on the total wheel forces and the vehicle speed and describes the energy temporarily stored in the vehicle. The total wheel forces are expressed as positive for the drive and negative for the brake. The temporarily stored energy can be embodied, for example, as the chemical energy contained in gasoline in an internal combustion engine and / or the state of charge in a battery.

[0021] Energy consumption and travel time are preferably evaluated and weighted at the end of the horizon. Therefore, both the efficiency term and the time term are only relevant at the last distance in the horizon. Therefore, the generalized efficiency term for the temporarily stored energy is approximated by the vehicle speed and wheel forces, or by the gradient of the temporarily stored energy as a function of the vehicle speed and the total wheel forces.

[0022] In contrast to the previous representation by means of specific engine torques and / or motor torques or their gradients, this has the advantage that the structure of the processor unit according to the invention is the same for all drive train variants and can therefore be used for a wide variety of drive train variants.

[0023] The processor unit according to the invention can therefore be used for all available drive trains, ie for internal combustion engines, hybrid vehicles, electric vehicles with a different number of electric motors, fuel cell vehicles, etc.

[0024] For the purpose of application, the processor unit can be used in any motor vehicle and, for example, transmit the determined speed to the power electronics.

[0025] Additionally, the efficiency term may be based on vehicle speed and wheel forces, on engine temperature, or on gradient.

[0026] In a further embodiment, the motor vehicle has a battery, wherein the efficiency term is based on the vehicle speed and the wheel forces as well as the engine temperature and the battery charge state. This allows the efficiency term to be reflected more accurately under different conditions.

[0027] In a further embodiment, the processor unit is configured to minimize the cost function using the gradient of the efficiency term, wherein the gradient is approximated using a polynomial, a neural network, a radial basis function, and / or a Gaussian process regression. In this case, the processor unit does not minimize the cost function using the gradient of the efficiency term itself, but rather uses the coefficients of the polynomial approximation. This simplifies and reduces the necessary information exchange with the processor unit.

[0028] In a further embodiment, the vehicle speed is formed according to the longitudinal dynamics model of the motor vehicle. and thus form the time term. Here, we can Forming a longitudinal dynamic model, in, represents the vehicle mass, and is the hill resistance and is dependent on the gradient of the road surface, which describes the longitudinal component of the weight force acting on the motor vehicle during uphill or downhill driving, represents the air resistance of the motor vehicle, wherein the air resistance depends on the speed of the motor vehicle; and is the rolling resistance, which is a result of the deformation of the tire as it rolls and depends on the wheel load and therefore on the slope angle of the road, and Represents the total wheel force.

[0029] In a further construction scheme, it is possible to Forming a longitudinal dynamic model, in, describes the kinetic energy of the vehicle, is the hill resistance and is dependent on the gradient of the road surface. It describes the longitudinal component of the weight acting on the motor vehicle during uphill or downhill driving. represents the air resistance of the motor vehicle (2), wherein the air resistance depends on the speed of the motor vehicle (2); and is the rolling resistance, which is a result of the deformation of the tire as it rolls and depends on the wheel load and therefore on the slope angle of the road, and Represents the wheel force.

[0030] This representation allows the longitudinal dynamics model to be applied to any desired drive train. Furthermore, it allows current state variables to be measured, corresponding data to be acquired, and these to be fed to the MPC module. Thus, for example, route data from an electronic map for the predicted horizon ahead of the motor vehicle can be regularly updated.

[0031] The route data may contain, for example, gradient information, curve information, and information about speed limits.

[0032] Therefore, vehicle parameters and the knowledge of the route topography ahead, such as bends and slopes, can be incorporated into this general longitudinal dynamics model. In addition, the knowledge of the speed limit on the route ahead can also be incorporated into the longitudinal dynamics model.

[0033] However, the general longitudinal dynamics model is not dependent on a specific drive train.

[0034] In a further construction scheme, the cost function has a comfort term weighted by a third weighting factor, which comfort term includes a predicted value of the wheel force change during the prediction horizon, wherein the processor unit is configured to: generate a speed control for the motor vehicle by implementing an MPC module in a time-dependent, efficiency-dependent and comfort-dependent manner so as to minimize the cost function.

[0035] To ensure comfortable driving, a comfort term is added to the cost function in order to ensure jerk-free driving.

[0036] The wheel forces at the operating point of the motor vehicle during the prediction horizon are represented by the difference : in, as the traction applied to the wheels of a motor vehicle by an engine or engines and As braking force.

[0037] In this case, the processor unit is configured to generate a speed control for the motor vehicle by executing an MPC module in a time-dependent, efficiency-dependent and comfort-dependent manner, such that a cost function is minimized.

[0038] Furthermore, the object is achieved by a method for predictive control of a motor vehicle, comprising the following steps: - Provides MPC modules with cost functions, wherein the cost function has a time term weighted by a first weighting factor, which term contains the travel time required by the motor vehicle, independently of the drive train variant of the motor vehicle, to cover the entire distance predicted within the prediction horizon, and - wherein the cost function has an efficiency term weighted by a second weighting factor, which efficiency term provides an energy consumption prediction within a prediction horizon that is dependent at least on the vehicle speed and the wheel forces, wherein the provided efficiency term is independent of a drive train variant of the motor vehicle, wherein the MCP module is implemented as a function of a weighted efficiency term and a weighted time term in order to determine a speed control for the motor vehicle over the prediction horizon independently of its drive train variant, such that a cost function is minimized.

[0039] The advantages of the processor unit can also be transferred to the method, which is designed in particular to be executed on the processor unit according to the invention.

[0040] Here, the vehicle speed can be formed according to the longitudinal dynamics model of the motor vehicle and thus form the time term, where, by Forming a longitudinal dynamic model, in, represents the vehicle mass, and is the hill resistance and is dependent on the gradient of the road surface. It describes the longitudinal component of the weight acting on the motor vehicle during uphill or downhill driving. represents the air resistance of the motor vehicle, wherein the air resistance depends on the speed of the motor vehicle; and is the rolling resistance, which is a result of the deformation of the tire as it rolls and depends on the wheel load and therefore on the slope angle of the road, and Represents the total wheel force.

[0041] Furthermore, the cost function has a comfort term weighted by a third weighting factor, which comfort term contains predicted values ​​of wheel force variations during the prediction horizon, and wherein, by implementing the MPC module in a time-dependent, efficiency-dependent and comfort-dependent manner, a speed control for the motor vehicle is generated such that the cost function is minimized.

[0042] Furthermore, the wheel forces may be formed by the difference between the traction and braking forces applied to the wheels of the motor vehicle by the engine or engines.

[0043] Furthermore, the object is achieved by a computer program product for predictive control of a motor vehicle, wherein, when the computer program product is executed on a processor unit, the computer program product causes the processor unit to execute an MPC module having a cost function, wherein the cost function has a time term weighted by a first weighting factor, which time term contains the travel time required by the motor vehicle for covering the entire distance predicted within the prediction horizon, predicted independently of a drive train variant of the motor vehicle, and wherein the cost function has an efficiency term weighted by a second weighting factor, which efficiency term provides an energy consumption prediction within a prediction horizon that depends at least on the vehicle speed and the wheel forces, wherein the provided efficiency term is independent of a drive train variant of the motor vehicle, and By implementing the MCP module as a function of a weighted efficiency term and a weighted time term, a speed control for the motor vehicle is generated for the motor vehicle over the prediction horizon independently of its drive train variant, minimizing a cost function.

[0044] Such a computer program product can be used in any motor vehicle with any drive train. In particular, due to its independence from the drive train, such a computer program product can be installed in any suitable motor vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Further features and advantages of the present invention will be apparent from the following description with reference to the accompanying drawings.

[0046] Figure 1 A processor unit for a motor vehicle is shown;

[0047] Figure 2 shows the gradient of the efficiency term for an electric motor vehicle;

[0048] Figure 3 The gradient of the efficiency term is shown for a motor vehicle of the internal combustion engine type. DETAILED DESCRIPTION

[0049] Figure 1 A processor unit 1 according to the invention is schematically shown for a motor vehicle 2. The processor unit 1 comprises a memory unit 4 and a detection unit, in particular a sensor system 3, for detecting status data relating to the motor vehicle 2 and, if appropriate, preceding route data.

[0050] Furthermore, the motor vehicle 2 comprises a drive train, which can be configured, for example, as an electric motor that can operate as a motor and as a generator, and an energy supply unit 6, which can be configured, for example, as a battery / gasoline tank / hydrogen tank for an internal combustion engine or a hybrid drive and / or a fuel cell.

[0051] In this case, the memory unit 4 has an MPC (Model Predictive Control) module 5 , which contains a cost function J to be minimized.

[0052] In this case, the MPC (Model Predictive Control) module 5 has a universal longitudinal dynamics model, which is applicable to all drive trains or any motor vehicle 2 :

[0053] Here, through Forming a general longitudinal dynamics model, in, describes the kinetic energy of the vehicle, is the hill resistance and is dependent on the gradient of the road surface, it describes the longitudinal component of the weight acting on the motor vehicle 2 during uphill or downhill driving, represents the air resistance of the motor vehicle 2 , wherein the air resistance depends on the speed of the motor vehicle 2 ; and is the rolling resistance, which is a result of the deformation of the tire as it rolls and depends on the wheel load and therefore on the slope angle of the road, and Represents the wheel force.

[0054] The time dependency is transformed into the distance dependency as follows:

[0055] By means of these general specifications, current state variables of the motor vehicle 2 can be measured and / or received, for example, by means of the sensor system 3 and supplied to the MPC module 5 .

[0056] Thus, for example, route data / surroundings data from an electronic map for a predicted horizon, for example, a few hundred meters ahead of motor vehicle 2, can be regularly updated. The surroundings data can include, for example, slope information, curve information, and information about speed limits. Furthermore, the curvature of the curve can be converted into the speed limit for motor vehicle 2 via the maximum permissible lateral acceleration. Furthermore, the position of motor vehicle 2 can be determined, in particular, using a GPS signal for accurate positioning on an electronic map.

[0057] Determine the first weighting factor with the help of the longitudinal dynamics model A weighted time term Z, which contains the driving time that the motor vehicle 2 needs to cover the entire distance predicted within the prediction horizon, independently of the drivetrain variant of the motor vehicle 2 .

[0058] The time term Z forms the first term of the cost function J as follows: ,in, in, is a tuning parameter used to compromise between time and efficiency (here energy consumption), and is the milepost at the end of the forecast horizon, and t represents the time, t Therefore, it indicates the mileage point The travel time at the time of arrival and Represents the weighting factor.

[0059] The cost function J thus contains the first weighting factor The first term is the travel time required by the motor vehicle 2 to cover the predicted distance, which is weighted and predicted according to the longitudinal dynamics model. This means that, depending on the selection of the weighting factors, lower speeds are not always evaluated as optimal, and therefore there is no longer the problem of the resulting speed always being at the lower limit of the permissible speed range.

[0060] By using this general longitudinal dynamics model, the time term Z does not depend on the drive train that happens to be used.

[0061] Furthermore, the cost function J has an efficiency term E weighted by a second weighting factor, which provides an efficiency term that depends at least on the driving speed within the prediction horizon. and wheel force The predicted energy consumption is used as temporarily stored energy, wherein the efficiency term E is provided independently of the drive train variant of the motor vehicle 2 : ,in, in, is a tuning parameter used to compromise between time and efficiency and is a weighting factor, and wherein the efficiency term E is incorporated into the cost function J in a negative manner.

[0062] Furthermore, the cost function J to be minimized has a third comfort term K: The cost function J = Z – E + K is thus formed as follows: in, is a tuning parameter used to compromise between time and efficiency and is a positive weighting factor.

[0063] The cost function J therefore contains an additional term as a comfort term K. It should be noted that this term depends only on the wheel force changes, that is, only on the mileage points Wheel force gradient in and no longer relies on longitudinal dynamics models.

[0064] Here, the wheel force can be used as traction applied by one or more engines to the wheels of the motor vehicle 2 and braking force The difference: : the tractive force applied to the wheels of a motor vehicle by an engine or engines, : Braking force.

[0065] In addition, the efficiency term E as the second term depends only on the vehicle speed and wheel force and no longer depends on the longitudinal dynamics model. Optionally, the efficiency term E can include additional variables such as the engine temperature T. Therefore, the gradient of the efficiency term E for minimizing the cost function J can be expressed as follows: in, : Engine temperature, : Wheel force : Vehicle speed. Among them, by transforming the time correlation into the distance correlation, the applicable

[0066] The efficiency term E can thus be calculated by the gradient of the energy stored in the motor vehicle 2 As wheel force and vehicle speed and optionally as a function of other variables such as temperature T. For example, if a battery is present, the battery state of charge (battery SoC (State of Charge)) can also be taken into account.

[0067] gradient It can be based on wheel power and vehicle speed It is determined taking into account the power balance and the loss profile and, if appropriate, the shifting strategy and hybrid strategy for the respective drive train.

[0068] However, modeling of the drive train to determine the efficiency term E is not mandatory.

[0069] By minimizing this cost function J, the MPC module 5 can provide the prediction horizon for the mileage points. For the use of the processor unit 1 in the motor vehicle 2 , a software module can be connected downstream of the processor unit 1 , which software module detects currently relevant states and communicates them to the power electronics.

[0070] The energy-optimal speed trajectory for the predicted forward horizon can be calculated online on the processor unit 1 , which can in particular form a component of a central control unit of the motor vehicle 2 .

[0071] By using the MPC module 5 and the universal efficiency term E, speed control of the motor vehicle 2 can be achieved without exact knowledge of the drive train or a modeled drive train.

[0072] Through the total wheel force or wheel force gradient and the energy temporarily stored in the motor vehicle 2 as a gasoline (internal combustion engine) tank and / or battery charge state (electric motor vehicle) Formulating the cost function J can eliminate the need for accurate understanding or modeling of the drive system.

[0073] In contrast to the previous formulation or modeling of specific engine and / or electric machine torques or their gradients, the cost function J can now be used for all drivetrain variants, and thus the processor unit 1 according to the present invention can be used for a wide variety of drivetrain variants. This can save both development effort and time.

[0074] Here, the energy consumption and travel time can be evaluated and weighted separately at the end of the forecast horizon. This term is therefore only valid for the last point of the forecast horizon. It works.

[0075] Efficiency information is provided here as Provided, it is generally defined as the gradient of energy stored in the motor vehicle 2 , for example in a battery or a fuel / hydrogen tank.

[0076] This depends at least on the wheel forces and vehicle speed and optionally also on other variables such as the engine temperature T, for example.

[0077] Figure 2 This efficiency information is shown as a graph for a drive train of a purely electric vehicle having an electric machine EM and two transmission stages G1, G2. .

[0078] Here, speed In m / s, and the wheel force In kN and with gradient The unit is kW.

[0079] Furthermore, the engine temperature T and the power balance as well as the loss profile can already be taken into account in this diagram, and if necessary, the shifting strategy and the hybrid strategy.

[0080] So, for example, in the value kN and speed = 5 m / s, a recovery and gradient of slightly less than 50 kW can be generated The corresponding value is positive.

[0081] In addition, for example = 10 m / s and the wheel force = 10 kN, the gradient is obtained The corresponding negative value can reflect the energy consumption under this driving condition.

[0082] Here, the gradient This can be done with simple functions, such as and wheel force Other basis functions and approximation techniques, such as neural networks, radial basis functions, or Gaussian process regression, are also possible. In this case, the efficiency term E is determined by the coefficients of the polynomial approximation.

[0083] The processor unit 1 is therefore not in this case dependent on the gradient of the efficiency term Instead of minimizing the cost function J using the comprehensive characteristic curve of , the coefficients of the polynomial approximation are used. This simplifies and reduces the required computational effort in the processor unit 1.

[0084] Figure 3 This efficiency information is shown as a graph for a drive train of a purely internal combustion engine vehicle with an internal combustion engine VM and six transmission stages G1, G2, G3, G4, G5, G6. .

[0085] Furthermore, the engine temperature T and the power balance as well as the loss profile can already be taken into account in this diagram, and if necessary the shifting strategy and the hybrid strategy.

[0086] So, for example, in the value kN and speed = 5 m / s. Since no recuperation occurs, the value zero is used as the corresponding value in the cost function J. .

[0087] In addition, =10 m / s and the wheel force = 10 kN, the corresponding negative value can be used as , the negative value reflects the energy consumption under this driving condition.

[0088] Here, the gradient It is also possible to use simple functions, such as and wheel force Other basis functions and approximation techniques are also possible. In this case, the efficiency term does not include the gradient itself, but rather, for example, includes the coefficients of a polynomial approximation.

[0089] Therefore, the processor unit 1 in this case does not rely on the gradient of the efficiency term The cost function J is not minimized by itself, but the coefficients of the polynomial approximation are used. This simplifies and reduces the required computational effort in the processor unit 1.

[0090] The processor unit according to the invention provides a generalized statement of efficiency using the efficiency term according to the invention, which uniformly provides the current energy consumption with respect to the forecast horizon for a wide variety of drivetrain variants.

[0091] In addition, a generalized comfort term is provided that is independent of the longitudinal dynamics model.

[0092] Furthermore, a generally applicable longitudinal dynamics model which is independent of the drive train and can be used for the time term Z is provided.

[0093] In contrast to the previous formulation via specific engine and / or electric machine torques or their gradients, the processor unit 1 can therefore be used for all drive train variants.

[0094] Reference Signs List

[0095] 1 processor unit

[0096] 2 Motor Vehicles

[0097] 3 Sensor system

[0098] 4 storage units

[0099] 5 MPC (Model Predictive Control) module

[0100] 6 Energy supply unit

[0101] G1, ..., G6 gear stages

[0102] EM motor

[0103] VM internal combustion engine

[0104] Z time term

[0105] K Comfort item

[0106] E efficiency term

[0107] J Cost Function

[0108] T Engine temperature

[0109] Vehicle speed

[0110] Wheel force

[0111] Wheel force variation or wheel force gradient

Claims

1. A processor unit (1) for predictively controlling a motor vehicle (2), wherein: The processor unit (1) has a storage unit (4) with an MPC (Model Predictive Control) module (5), the MPC module including a cost function (J), Wherein, the cost function (J) has a first weighting factor ( ) a weighted time term (Z), which contains the travel time required by the motor vehicle (2) to cover the entire distance predicted within the prediction horizon, independently of the drive train variant of the motor vehicle (2), and Wherein, the cost function (J) has a second weighting factor ( ) weighted efficiency term (E), which provides a prediction horizon that depends at least on the vehicle speed ( ) and wheel force ( ) predicted energy consumption, wherein the efficiency term (E) provided is independent of the drive train variant of the motor vehicle (2), The processor unit (1) is configured to: determine a speed control for the motor vehicle (2) with respect to the prediction horizon independently of a drive train variant of the motor vehicle (2) by implementing an MCP module (5) in dependence on a weighted efficiency term (E) and a weighted time term (Z) so that the cost function (J) is minimized.

2. The processor unit (1) according to claim 1, characterized in that The efficiency term (E) is based on the vehicle speed ( ) and the wheel force ( ) and engine temperature (T).

3. The processor unit (1) according to claim 2, characterized in that The motor vehicle (2) has a battery and the efficiency term (E) is based on the vehicle speed ( ), the wheel force ( ), the engine temperature (T) and the battery charge state.

4. The processor unit (1) according to any one of the preceding claims 2 to 3, characterized in that The processor unit (1) is configured to minimize the cost function (J) using the gradient of the efficiency term (E), wherein the gradient is approximated by a polynomial, a neural network, a radial basis function and / or a Gaussian process regression.

5. Processor unit (1) according to any one of the preceding claims, characterized in that The time term (Z) is formed according to a longitudinal dynamics model of the motor vehicle (2).

6. The processor unit (1) according to claim 5, characterized in that The longitudinal dynamics model is form, in, describes the kinetic energy of the vehicle, is the hill resistance and is dependent on the gradient of the road surface, which describes the longitudinal component of the weight acting on the motor vehicle (2) during uphill or downhill driving, represents the air resistance of the motor vehicle (2), wherein the air resistance depends on the speed of the motor vehicle (2); and is the rolling resistance, which is a result of the deformation of the tire while rolling and depends on the load on the wheel and therefore on the slope angle of the road, and Represents the wheel force.

7. Processor unit (1) according to any one of the preceding claims, characterized in that The cost function (J) has a third weighting factor ( ) weighted comfort term (K), which includes the wheel force changes during the prediction horizon ( ), and wherein the processor unit (1) is configured to generate a speed control for the motor vehicle (2) by implementing an MPC module (5) in dependence on the time term (Z), in dependence on the efficiency term (E) and in dependence on the comfort term (K) so as to minimize the cost function (J).

8. Processor unit (1) according to any one of the preceding claims, characterized in that The wheel force ( ) is caused by the traction force applied to the wheels of the motor vehicle (2) by an engine or engines ( ) and braking force ( ) is formed by the difference.

9. The processor unit (1) according to claim 8, characterized in that The wheel force variation about the prediction horizon is formed ( ).

10. A method for predictively controlling a motor vehicle (2), the method comprising the steps of: - Provides an MPC module with a cost function (5), - in, The cost function (J) has a first weighting factor ( ) a weighted time term (Z), which contains the travel time required by the motor vehicle (2) to cover the entire distance predicted within the prediction horizon, independently of the drive train variant of the motor vehicle (2), and Wherein, the cost function (J) has a second weighting factor ( ) weighted efficiency term (E), which provides a dependence of at least the vehicle speed ( ) and wheel force ( ) predicted energy consumption, wherein the efficiency term (E) provided is independent of the drive train variant of the motor vehicle (2), - implementing an MCP module (5) as a function of a weighted efficiency term (E) and a weighted time term (Z) in order to determine a speed control for the motor vehicle (2) with respect to the prediction horizon independently of the drive train variant of the motor vehicle (2) such that the cost function (J) is minimized.

11. The method according to claim 10, characterized in that The time term (Z) is formed according to a longitudinal dynamics model of the motor vehicle (2), wherein forming the longitudinal dynamic model, in, describes the kinetic energy of the vehicle, is the hill resistance and is dependent on the gradient of the road surface, which describes the longitudinal component of the weight acting on the motor vehicle (2) during uphill or downhill driving, represents the air resistance of the motor vehicle (2), wherein the air resistance depends on the speed of the motor vehicle (2); and is the rolling resistance, which is a result of the deformation of the tire while rolling and depends on the load on the wheel and therefore on the slope angle of the road, and Represents the wheel force.

12. The method according to claim 10 or 11, characterized in that The cost function (J) has a third weighting factor ( ) weighted comfort term (K), which includes the wheel force changes during the prediction horizon ( ), and wherein, by implementing an MPC module (5) in dependence on the time term (Z), in dependence on the efficiency term (E) and in dependence on the comfort term (K), a speed control for the motor vehicle (2) is generated such that the cost function (J) is minimized.

13. The method according to claim 12, characterized in that The wheel force ( ) is caused by the traction force applied to the wheels of the motor vehicle (2) by an engine or engines ( ) and braking force ( ) is formed by the difference.

14. A computer program product for predictively controlling a motor vehicle (2), wherein: When the computer program product is executed on a processor unit (1), the computer program product causes the processor unit (1) to execute an MPC module (5) having a cost function (J) stored in a memory unit (4), wherein the cost function (J) has a weighting factor ( ) weighted time term (Z), which contains the travel time required by the motor vehicle (2) to cover the entire distance predicted within the prediction horizon, independently of the drive train variant of the motor vehicle (2), and wherein the cost function (J) has a weighted time term ( ) weighted efficiency term (E), which provides a prediction horizon that depends at least on the vehicle speed ( ) and wheel force ( ) predicted energy consumption, wherein the efficiency term (E) provided is independent of the drive train variant of the motor vehicle (2), and wherein, By implementing an MCP module (5) as a function of a weighted efficiency term (E) and a weighted time term (Z), a speed control for the motor vehicle (2) with respect to the prediction horizon is generated for the motor vehicle (2) independently of the drive train variant of the motor vehicle (2), such that the cost function (J) is minimized.

Images